The n e w e ng l a n d j o u r na l of m e dic i n e

Original Article

Cardiovascular and Renal Outcomes

with Empagliflozin in Heart Failure
M. Packer, S.D. Anker, J. Butler, G. Filippatos, S.J. Pocock, P. Carson, J. Januzzi,
S. Verma, H. Tsutsui, M. Brueckmann, W. Jamal, K. Kimura, J. Schnee, C. Zeller,
D. Cotton, E. Bocchi, M. Böhm, D.-J. Choi, V. Chopra, E. Chuquiure, N. Giannetti,
S. Janssens, J. Zhang, J.R. Gonzalez Juanatey, S. Kaul, H.-P. Brunner‑La Rocca,
B. Merkely, S.J. Nicholls, S. Perrone, I. Pina, P. Ponikowski, N. Sattar, M. Senni,
M.-F. Seronde, J. Spinar, I. Squire, S. Taddei, C. Wanner, and F. Zannad,
for the EMPEROR-Reduced Trial Investigators*​​

A BS T R AC T

BACKGROUND
Sodium–glucose cotransporter 2 (SGLT2) inhibitors reduce the risk of hospitaliza- The authors’ full names, academic de-
tion for heart failure in patients regardless of the presence or absence of diabetes. grees, and affiliations are listed in the
Appendix. Address reprint requests to
More evidence is needed regarding the effects of these drugs in patients across the Dr. Packer at Baylor Heart and Vascular
broad spectrum of heart failure, including those with a markedly reduced ejection Institute, 621 N. Hall St., Dallas, TX
fraction. 75226, or at ­milton​.­packer@​­baylorhealth​
.­edu.
METHODS *A complete list of the EMPEROR-Reduced
In this double-blind trial, we randomly assigned 3730 patients with class II, III, or investigators is provided in the Supple-
mentary Appendix, available at NEJM
IV heart failure and an ejection fraction of 40% or less to receive empagliflozin .org.
(10 mg once daily) or placebo, in addition to recommended therapy. The primary
This article was published on August 29,
outcome was a composite of cardiovascular death or hospitalization for worsening 2020, at NEJM.org.
heart failure.
DOI: 10.1056/NEJMoa2022190
Copyright © 2020 Massachusetts Medical Society.
RESULTS
During a median of 16 months, a primary outcome event occurred in 361 of 1863
patients (19.4%) in the empagliflozin group and in 462 of 1867 patients (24.7%)
in the placebo group (hazard ratio for cardiovascular death or hospitalization for
heart failure, 0.75; 95% confidence interval [CI], 0.65 to 0.86; P<0.001). The effect
of empagliflozin on the primary outcome was consistent in patients regardless of
the presence or absence of diabetes. The total number of hospitalizations for heart
failure was lower in the empagliflozin group than in the placebo group (hazard
ratio, 0.70; 95% CI, 0.58 to 0.85; P<0.001). The annual rate of decline in the esti-
mated glomerular filtration rate was slower in the empagliflozin group than in
the placebo group (–0.55 vs. –2.28 ml per minute per 1.73 m2 of body-surface area
per year, P<0.001), and empagliflozin-treated patients had a lower risk of serious
renal outcomes. Uncomplicated genital tract infection was reported more frequently
with empagliflozin.
CONCLUSIONS
Among patients receiving recommended therapy for heart failure, those in the em-
pagliflozin group had a lower risk of cardiovascular death or hospitalization for
heart failure than those in the placebo group, regardless of the presence or absence
of diabetes. (Funded by Boehringer Ingelheim and Eli Lilly; EMPEROR-Reduced
ClinicalTrials.gov number, NCT03057977.)

n engl j med  nejm.org 1

The New England Journal of Medicine
Downloaded from nejm.org by CARLOS ANDRES TEJEDA PEREZ on October 2, 2020. For personal use only. No other uses without permission.
Copyright © 2020 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

I
n patients with type 2 diabetes, so- The trial was performed at 520 centers in 20 coun-
dium–glucose cotransporter 2 (SGLT2) in- tries. The protocol and the statistical analysis
hibitors reduce the risk of hospitalization for plan are available (as a single PDF file) with the
heart failure and the risk of serious adverse renal full text of this article at NEJM.org. The ethics
events, benefits that are not seen with other committee at each trial center approved the trial,
antihyperglycemic drugs. In large-scale, random- and all the patients provided written informed
ized, placebo-controlled trials, the risk of hospi- consent. The sponsors were Boehringer Ingelheim
talization for heart failure was 30 to 35% lower and Eli Lilly.
among patients who received SGLT2 inhibitors The executive committee developed and amend-
than among those who received placebo1; this ed the protocol and had scientific oversight on the
benefit was most striking in patients who had a development of the statistical analysis plan, the
left ventricular ejection fraction of 30% or less case report forms, the recruitment of patients,
before treatment.2 In addition, the risk of pro- the quality and thoroughness of follow-up, and
gression of renal disease (including the occurrence the analysis of data. The academic members of the
of renal death or the need for dialysis or renal executive committee provided an independent in-
transplantation) was 35 to 50% lower among pa- terpretation of the results. An independent data
tients who received SGLT2 inhibitors than among and safety monitoring committee reviewed the
those who received placebo.1 These cardiorenal safety data and the results of an interim analysis
benefits cannot be explained by an action of according to prespecified stopping boundaries.
SGLT2 inhibitors to lower blood glucose, since The statistical analyses were performed by em-
similar effects have not been seen with other ployees of the sponsor with the oversight of the
antidiabetic drugs that have greater antihyper- academic trial leadership, and an independent
glycemic actions.3 statistician replicated and verified the analyses.
These observations are consistent with the The first author, who had unrestricted access to
hypothesis that SGLT2 inhibitors may slow the the data, prepared the first draft of the manu-
progression of cardiac and renal disease, regard- script, which was then reviewed and edited by all
less of cause and independent of the presence or the authors. The authors made the decision to
absence of diabetes.3 The Dapagliflozin and Pre- submit the manuscript for publication, assume
vention of Adverse Outcomes in Heart Failure full responsibility for the accuracy and complete-
(DAPA-HF) trial showed a reduction in the risk ness of the analyses, and attest to the fidelity of
of cardiovascular death or hospitalization for the trial to the protocol and the statistical analy-
heart failure with dapagliflozin in patients re- sis plan.
gardless of the presence or absence of diabetes4;
this trial primarily enrolled patients with mild- Patients
to-moderate degrees of left ventricular systolic Adults (≥18 years of age) who had chronic heart
dysfunction and increases in natriuretic peptide failure (functional class II, III, or IV) with a left
levels. In the Empagliflozin Outcome Trial in ventricular ejection fraction of 40% or less were
Patients with Chronic Heart Failure and a Reduced eligible to participate in the trial. All the patients
Ejection Fraction (EMPEROR-Reduced), we eval- were receiving appropriate treatments for heart
uated empagliflozin in a population of patients failure, including diuretics, inhibitors of the
with chronic heart failure and a reduced ejection renin–angiotensin system and neprilysin, beta-
fraction (with or without diabetes) that was en- blockers, mineralocorticoid receptor antagonists,
riched for patients with a greater severity of leftand, when indicated, cardiac devices.
ventricular systolic dysfunction. The intent of the trial was to enroll patients
with heart failure who were at increased risk for
a serious heart failure event. We limited the num-
Me thods
ber of patients who had an ejection fraction of
Trial Design and Oversight more than 30% by requiring a history of hospi-
Details regarding the design of this randomized, talization for heart failure within the previous
double-blind, parallel-group, placebo-controlled, 12 months or a particularly high level of N-ter-
event-driven trial have been reported previously.5 minal prohormone of brain natriuretic peptide

2 n engl j med  nejm.org

The New England Journal of Medicine

Downloaded from nejm.org by CARLOS ANDRES TEJEDA PEREZ on October 2, 2020. For personal use only. No other uses without permission.
Copyright © 2020 Massachusetts Medical Society. All rights reserved.
 Cardiovascular and Renal Outcomes with Empagliflozin
(NT-proBNP), including a level of at least 1000 pg
per milliliter in those with an ejection fraction
of 31 to 35% or a level of at least 2500 pg per
milliliter in those with an ejection fraction of
36 to 40%, as compared with a level of at least
600 pg per milliliter in those with an ejection
fraction of 30% or less.6 These NT-proBNP thresh-
olds were doubled in patients with atrial fibrilla-
tion.5 The key inclusion and exclusion criteria are
provided in the Supplementary Appendix, avail-
able at NEJM.org.
Trial Visits and Follow-up
After a screening period of 4 to 28 days, patients
who fulfilled the eligibility criteria were randomly
assigned in a 1:1 ratio to receive either empa-
gliflozin (at a dose of 10 mg daily) or placebo in
addition to their usual therapy for heart failure.
The dose of empagliflozin was selected on the
basis of the reduction in the risk of cardiovascu-
lar death or hospitalization for heart failure that
had been previously reported with this dose in
patients with type 2 diabetes.7 Randomization
was performed with an interactive-response sys-
tem that used a permuted-block design and was
stratified according to geographical region (North
America, Latin America, Europe, Asia, or other),
diabetes status at screening, and the estimated
glomerular filtration rate (GFR) at screening
(<60 or ≥60 ml per minute per 1.73 m2 of body-
surface area), according to the Chronic Kidney
Disease Epidemiology Collaboration (CKD-EPI)
equation. After randomization, all appropriate
treatments for heart failure or other medical con-
ditions could be initiated or altered at the clini-
cal discretion of the health care provider, accord-
ing to each patient’s needs.
Every 2 to 3 months, we evaluated patients at
trial visits to assess outcomes and adverse events.
We periodically evaluated vital signs, body weight,
glycated hemoglobin level, NT-proBNP level, and
renal function. In addition, we assessed the pa-
tients’ quality of life using the Kansas City Car-
diomyopathy Questionnaire. We reevaluated the
estimated GFR 23 to 45 days after the discon-
tinuation of empagliflozin or placebo in order to
allow for an evaluation of the effect of treatment
independent of the presence of the SGLT2 in-
hibitor. All the patients were followed for the
occurrence of prespecified outcomes for the
entire duration of the trial, regardless of wheth-
n engl j med  nejm.org
The New England Journal of Medicine
Downloaded from nejm.org by CARLOS ANDRES TEJEDA PEREZ on October 2, 2020. For personal use only. No other uses without permission.
Copyright © 2020 Massachusetts Medical Society. All rights reserved.
er they were adherent to the trial regimens or
procedures.
Primary and Secondary Outcomes
The primary outcome and the first two secondary
outcomes were included in a hierarchical testing
procedure, as described in the Statistical Analy-
sis section. The primary outcome was a compos-
ite of adjudicated cardiovascular death or hospi-
talization for heart failure, analyzed as the time
to the first event. The first secondary outcome
was the occurrence of all adjudicated hospital-
izations for heart failure, including first and re-
current events. The second secondary outcome
was the rate of the decline in the estimated GFR
during double-blind treatment.
Additional prespecified efficacy outcomes that
were not part of the testing hierarchy (including
a composite renal outcome, total hospitalizations
for any reason, and quality of life) are described
in the Supplementary Appendix. Safety analyses
included all the patients who had received at least
one dose of empagliflozin or placebo. A clinical-
events committee adjudicated fatal and nonfatal
events in a blinded manner according to pre-
specified definitions, which are provided in the
Supplementary Appendix.
Statistical Analysis
We determined that a target number of 841 ad-
judicated primary outcome events would provide
a power of 90% to detect a 20% lower relative
risk of the primary outcome in the empagliflozin
group than in the placebo group at a two-sided
alpha level of 0.05. Assuming an annual incidence
of the primary outcome of at least 15% per year in
the placebo group and a recruitment period of
18 months, we established a planned enrollment
of 2850 patients, with the option of increasing
the enrollment to 4000 patients if the accumula-
tion of primary outcome events was slower than
expected. Accordingly, the number of patients
undergoing randomization was increased to 3600
with no change to the target number of events.
This increase in sample size was made without
any knowledge of unblinded trial data and be-
fore the planned formal interim analysis of effi-
cacy by the data and safety monitoring commit-
tee. This committee carried out one prespecified
interim efficacy analysis after the occurrence of
approximately 500 primary outcome events, with
3
 The n e w e ng l a n d j o u r na l of m e dic i n e

Table 1. Characteristics of the Patients at Baseline.*

Empagliflozin Placebo
Characteristic (N = 1863) (N = 1867)
Age — yr 67.2±10.8 66.5±11.2
Female sex — no. (%) 437 (23.5) 456 (24.4)
Race — no. (%)†
White 1325 (71.1) 1304 (69.8)
Black 123 (6.6) 134 (7.2)
Asian 337 (18.1) 335 (17.9)
Other or missing 78 (4.2) 94 (5.0)
Region — no. (%)
North America 212 (11.4) 213 (11.4)
Latin America 641 (34.4) 645 (34.5)
Europe 676 (36.3) 677 (36.3)
Asia 248 (13.3) 245 (13.1)
Other 86 (4.6) 87 (4.7)
NYHA functional class — no. (%)
II 1399 (75.1) 1401 (75.0)
III 455 (24.4) 455 (24.4)
IV 9 (0.5) 11 (0.6)
Body-mass index‡ 28.0±5.5 27.8±5.3
Heart rate — beats/min 71.0±11.7 71.5±11.8
Systolic blood pressure — mm Hg 122.6±15.9 121.4±15.4
Left ventricular ejection fraction
Mean value 27.7±6.0 27.2±6.1
Value of ≤30% — no. (%) 1337 (71.8) 1392 (74.6)
NT-proBNP
Median value (IQR) — pg/ml 1887 (1077–3429) 1926 (1153–3525)
Value of ≥1000 pg/ml — no./total no. (%) 1463/1862 (78.6) 1488/1866 (79.7)
Cause of heart failure — no. (%)
Ischemic 983 (52.8) 946 (50.7)
Nonischemic 880 (47.2) 921 (49.3)
Cardiovascular history — no. (%)
Hospitalization for heart failure in ≤12 mo 577 (31.0) 574 (30.7)
Atrial fibrillation 664 (35.6) 705 (37.8)
Diabetes mellitus 927 (49.8) 929 (49.8)
Hypertension 1349 (72.4) 1349 (72.3)
Estimated glomerular filtration rate
Mean value — ml/min/1.73 m2 61.8±21.7 62.2±21.5
Value of <60 ml/min/1.73 m2 — no./total no. (%) 893/1862 (48.0) 906/1866 (48.6)

4 n engl j med  nejm.org

The New England Journal of Medicine

Downloaded from nejm.org by CARLOS ANDRES TEJEDA PEREZ on October 2, 2020. For personal use only. No other uses without permission.
Copyright © 2020 Massachusetts Medical Society. All rights reserved.
 Cardiovascular and Renal Outcomes with Empagliflozin

Table 1. (Continued)

Empagliflozin Placebo
Characteristic (N = 1863) (N = 1867)
Heart failure medication — no. (%)
Renin–angiotensin inhibitor§
Without neprilysin inhibitor 1314 (70.5) 1286 (68.9)
With neprilysin inhibitor 340 (18.3) 387 (20.7)
Mineralocorticoid receptor antagonist 1306 (70.1) 1355 (72.6)
Beta-blocker 1765 (94.7) 1768 (94.7)
Device therapy — no. (%)
Implantable cardioverter–defibrillator¶ 578 (31.0) 593 (31.8)
Cardiac resynchronization therapy‖ 220 (11.8) 222 (11.9)

* Plus–minus values are means ±SD. Percentages may not total 100 because of rounding. IQR denotes interquartile
range, NT-proBNP N-terminal prohormone of brain natriuretic peptide, and NYHA New York Heart Association.
† Race was reported by the patients. Those who identified with more than one race or with no race were classified as
“other.”
‡ The body-mass index is the weight in kilograms divided by the square of the height in meters.
§ Inhibitors of the renin–angiotensin system include angiotensin-converting–enzyme inhibitors and angiotensin-receptor
blockers.
¶ This category includes all the patients with an implantable cardioverter–defibrillator regardless of the presence or ab-
sence of cardiac resynchronization therapy.
‖ This category includes all the patients who were receiving cardiac resynchronization therapy regardless of the presence
or absence of a defibrillator.

the possibility of recommending early termination
of the trial if a benefit associated with empa-
gliflozin was significant at a one-sided alpha level
of approximately 0.001 with respect to both the
primary outcome and cardiovascular death alone.
The primary analysis was performed accord-
ing to the intention-to-treat principle and includ-
ed all the data that had been obtained up to the
end of the planned treatment period for all the
patients who had undergone randomization. Be-
tween-group differences in the primary outcome
were assessed for statistical significance with
the use of a Cox proportional-hazards model, with
prespecified covariates of age, sex, geographical
region, diabetes status at baseline, left ventricular
ejection fraction, and estimated GFR at baseline.
If the between-group difference in the primary
outcome was significant, the two key secondary
outcomes were prespecified to be analyzed in a
stepwise hierarchical manner to preserve the over-
all type I error rate at 0.0496 (two-sided) after
accounting for one interim analysis. The first sec-
ondary outcome — total (first and recurrent)
hospitalizations for heart failure — was evaluat-
ed with the use of a joint frailty model that ac-
counted for informative censoring because of
cardiovascular death. The second secondary out-
come (the slope of the change in the estimated
GFR) was evaluated at an alpha level of 0.001.
The remaining alpha level after hierarchical test-
ing will be applied to a patient-level meta-analy-
sis that will be performed on the data sets from
the current trial and from a parallel trial, Empa-
gliflozin Outcome Trial in Patients with Chronic
Heart Failure and a Preserved Ejection Fraction
(EMPEROR-Preserved).
R e sult s
Patients
From April 2017 through November 2019, a total
of 7220 patients were screened for eligibility,
and 3730 were randomly assigned to receive ei-
ther empagliflozin (1863 patients) or placebo
(1867 patients) (Fig. S1 in the Supplementary
Appendix). The reasons for screening failure are

n engl j med  nejm.org 5

The New England Journal of Medicine
Downloaded from nejm.org by CARLOS ANDRES TEJEDA PEREZ on October 2, 2020. For personal use only. No other uses without permission.
Copyright © 2020 Massachusetts Medical Society. All rights reserved.
 6
Table 2. Primary and Secondary Cardiovascular Outcomes.*

Hazard Ratio or Absolute

Variable Empagliflozin (N = 1863) Placebo (N = 1867) Difference (95% CI)† P Value

events/100 events/100
patient-yr patient-yr
Primary composite outcome — no. (%) 361 (19.4) 15.8 462 (24.7) 21.0 0.75 (0.65 to 0.86) <0.001
Hospitalization for heart failure 246 (13.2) 10.7 342 (18.3) 15.5 0.69 (0.59 to 0.81)
Cardiovascular death 187 (10.0) 7.6 202 (10.8) 8.1 0.92 (0.75 to 1.12)
Secondary outcomes specified in hierarchical testing
procedure
Total no. of hospitalizations for heart failure 388 — 553 — 0.70 (0.58 to 0.85) <0.001
Mean slope of change in eGFR — ml/min/1.73 m2 –0.55±0.23 — –2.28±0.23 — 1.73 (1.10 to 2.37) <0.001
The

per year‡
Other prespecified analyses
Composite renal outcome — no. (%)§ 30 (1.6) 1.6 58 (3.1) 3.1 0.50 (0.32 to 0.77)
Change in quality-of-life score on KCCQ at 52 weeks¶ 5.8±0.4 — 4.1±0.4 — 1.7 (0.5 to 3.0)
No. of hospitalizations for any cause 1364 — 1570 — 0.85 (0.75 to 0.95)
Death from any cause — no. (%) 249 (13.4) 10.1 266 (14.2) 10.7 0.92 (0.77 to 1.10)
Onset of new diabetes in patients with prediabetes 71/632 (11.2) 9.3 80/636 (12.6) 10.6 0.86 (0.62 to 1.19)
— no./total no. (%)
Laboratory and other measurements (adjusted change
from baseline to 52 wk)
Glycated hemoglobin in patients with diabetes — % –0.28±0.03 — –0.12±0.03 — –0.16 (–0.25 to –0.08)
n e w e ng l a n d j o u r na l

The New England Journal of Medicine

n engl j med  nejm.org
Hematocrit (%) 1.98±0.10 — –0.38±0.10 — 2.36 (2.08 to 2.63)
of

Median NT-proBNP (IQR) — pg/ml‖ –244 (–890 to 260) –141 (–784 to 585) — 0.87 (0.82 to 0.93)
Body weight — kg –0.73±0.13 — 0.08±0.13 — –0.82 (–1.18 to –0.45)
Systolic blood pressure — mm Hg –2.4±0.4 — –1.7±0.4 — –0.7 (–1.8 to 0.4)

Copyright © 2020 Massachusetts Medical Society. All rights reserved.

m e dic i n e

* Plus–minus values are means ±SE. IQR denotes interquartile range.

† All treatment effects are shown as hazard ratios, except for the slope of the change in the estimated glomerular filtration rate (GFR), the quality-of-life score, and the variables listed
under laboratory and other measurements. For all hazard ratios or treatment differences without P values, the widths of the confidence intervals have not been adjusted for multiple
comparisons, so the intervals should not be used to infer definitive treatment effects.
‡ The slope of the estimated GFR was analyzed on the basis of on-treatment data with a random coefficient model that included age and baseline estimated GFR as linear covariates and
sex, region, baseline left ventricular ejection fraction, baseline diabetes status, baseline estimated GFR according to time, and treatment according to time interactions as fixed effects;
the model allows for randomly varying slope and intercept between patients.
§ The composite renal outcome includes chronic dialysis or renal transplantation or a sustained reduction of 40% or more in the estimated GFR or a sustained estimated GFR of less
than 15 ml per minute per 1.73 m2 in patients with a baseline estimated GFR of 30 ml per minute per 1.73 m2 or more or a sustained estimated GFR of less than 10 ml per minute per
1.73 m2 in those with a baseline estimated GFR of less than 30 ml per minute per 1.73 m2.
¶ The clinical summary score on the Kansas City Cardiomyopathy Questionnaire (KCCQ) ranges from 0 to 100, with higher scores indicating a better quality of life.
‖ NT-proBNP was analyzed with the use of a geometric mean ratio because modeling was performed on log-transformed data.

Downloaded from nejm.org by CARLOS ANDRES TEJEDA PEREZ on October 2, 2020. For personal use only. No other uses without permission.
 Cardiovascular and Renal Outcomes with Empagliflozin
described in Table S1. The baseline characteris-
tics of the patients in the two trial groups were
similar (Table 1). Half the patients had a history
of diabetes, 73% had a left ventricular ejection
fraction of 30% or less, 79% had a NT-proBNP
level of at least 1000 pg per milliliter, 48% had
an estimated GFR of less than 60 ml per minute
per 1.73 m2, and nearly 20% were receiving an
angiotensin receptor–neprilysin inhibitor.
The final date of follow-up data collection for
the double-blind treatment period was April 29,
2020. Four patients in the placebo group did not
receive placebo. In addition, empagliflozin or pla-
cebo was stopped prematurely for reasons other
than death in 303 patients (16.3%) in the empa-
gliflozin group and in 335 patients (18.0%) in
the placebo group. A total of 21 patients (0.6%)
had unknown vital status at the end of the trial,
in part related to operational challenges associ-
ated with coronavirus disease 2019 (Covid-19);
42 patients (20 in the placebo group and 22 in
the empagliflozin group), including those with
unknown vital status, were lost to follow-up at
various times before the data cutoff. The median
duration of follow-up was 16 months.
Primary Outcome
The primary composite outcome of death from
cardiovascular causes or hospitalization for heart
failure occurred in 361 patients (19.4%) in the
empagliflozin group and in 462 patients (24.7%)
in the placebo group (hazard ratio, 0.75; 95%
confidence interval [CI], 0.65 to 0.86; P<0.001)
(Table 2 and Fig. 1A). The hazard ratios for the
effect of empagliflozin on cardiovascular death
and on the first hospitalization for heart failure
were 0.92 (95% CI, 0.75 to 1.12) and 0.69 (95% CI,
0.59 to 0.81), respectively (Table 2 and Figs. S2
and S3). During the trial period, the number of
patients who would need to have been treated with
empagliflozin to prevent one primary event was
19 (95% CI, 13 to 37).
The effect of empagliflozin on the primary
outcome was consistent across prespecified sub-
groups, including patients with diabetes and those
without diabetes at baseline (Fig. 2 and Fig. S4).
Among the patients who were receiving sacubi-
tril–valsartan at baseline, the hazard ratio for
the comparison between empagliflozin and pla-
cebo for the primary outcome was 0.64 (95% CI,
0.45 to 0.89), as compared with 0.77 (95% CI,
n engl j med  nejm.org
The New England Journal of Medicine
Downloaded from nejm.org by CARLOS ANDRES TEJEDA PEREZ on October 2, 2020. For personal use only. No other uses without permission.
Copyright © 2020 Massachusetts Medical Society. All rights reserved.
0.66 to 0.90) among those who were not receiv-
ing sacubitril–valsartan.
Secondary Outcomes
Empagliflozin also favorably influenced the two
prespecified secondary outcomes that were in-
cluded in the hierarchical testing procedure. The
total number of hospitalizations for heart failure
was lower in the empagliflozin group than in the
placebo group, with 388 events and 553 events,
respectively (hazard ratio, 0.70; 95% CI, 0.58 to
0.85; P<0.001) (Table 2 and Fig. 1B). The rate of
the decline in the estimated GFR over the dura-
tion of the double-blind treatment period was
slower in the empagliflozin group than in the
placebo group (–0.55 ml per minute per 1.73 m2
per year vs. –2.28 ml per minute per 1.73 m2 per
year), for a between-group difference of 1.73 ml
per minute per 1.73 m2 per year (95% CI, 1.10 to
2.37; P<0.001) (Table 2 and Fig. 3).
Other Prespecified Outcomes
In prespecified analyses that were not included
in the testing hierarchy, a composite renal out-
come (chronic dialysis or renal transplantation or
a profound, sustained reduction in the estimated
GFR) occurred in 30 patients (1.6%) in the em-
pagliflozin group and in 58 patients (3.1%) in
the placebo group (hazard ratio, 0.50; 95% CI,
0.32 to 0.77). In 966 patients with paired mea-
surements before treatment and 23 to 45 days
after the discontinuation of the trial regimens
(which enabled assessment of the effects of em-
pagliflozin independent of the presence of the
drug), the estimated GFR declined by –0.93 ml
per minute per 1.73 m2 (95% CI, –1.97 to 0.11)
in the empagliflozin group and by –4.21 ml per
minute per 1.73 m2 (95% CI, –5.26 to –3.17) in
the placebo group. The effects of empagliflozin
on patients’ quality of life, total hospitalizations
for any reason, and the frequency of new-onset
diabetes are described in Table 2. A total of 249
patients (13.4%) in the empagliflozin group and
266 patients (14.2%) in the placebo group died
from any cause (hazard ratio, 0.92; 95% CI, 0.77
to 1.10) (Fig. S5).
Biomarkers, Safety, and Sensitivity Analyses
Changes from baseline to 52 weeks in values for
glycated hemoglobin, hematocrit, NT-proBNP,
body weight, and systolic blood pressure in the
7
 The n e w e ng l a n d j o u r na l of m e dic i n e

A Primary Outcome
100 35
Hazard ratio, 0.75 (95% CI, 0.65–0.86) Placebo
P<0.001
90 30
Empagliflozin
25
80

Estimated Cumulative Incidence (%)

20
70
15
60
10

50 5

40 0
0 90 180 270 360 450 540 630 720 810
Placebo
30
Empagliflozin

20

10

0
0 90 180 270 360 450 540 630 720 810
Days since Randomization
No. at Risk
Placebo 1867 1715 1612 1345 1108 854 611 410 224 109
Empagliflozin 1863 1763 1677 1424 1172 909 645 423 231 101

B First and Recurrent Hospitalizations for Heart Failure

0.6
Hazard ratio, 0.70 (95% CI, 0.58–0.85)
P<0.001
0.5
Mean Number of Events per Patient

Placebo

0.4
Empagliflozin

0.3

0.2

0.1

0.0
0 90 180 270 360 450 540 630 720 810
Days since Randomization
No. at Risk
Placebo 1867 1820 1762 1526 1285 1017 732 497 275 135
Empagliflozin 1863 1826 1768 1532 1283 1008 732 495 272 118

two groups are shown in Table 2. The 4 patients frequently with empagliflozin than with placebo.
in the placebo group who did not receive placebo Adverse events of interest are listed in Table S2.
were excluded from the safety analyses. Uncom- Several sensitivity analyses were performed to
plicated genital tract infection was reported more account for missing follow-up data in 42 patients

8 n engl j med  nejm.org

The New England Journal of Medicine

Downloaded from nejm.org by CARLOS ANDRES TEJEDA PEREZ on October 2, 2020. For personal use only. No other uses without permission.
Copyright © 2020 Massachusetts Medical Society. All rights reserved.
 Cardiovascular and Renal Outcomes with Empagliflozin
Figure 1 (facing page). Primary Outcome and Total
Hospitalizations for Heart Failure.
Shown is the cumulative incidence of the primary com-
posite outcome of cardiovascular death or hospitaliza-
tion for heart failure (Panel A) and the total (first and
recurrent) hospitalizations for heart failure, expressed
as the mean number per patient (Panel B) in the em-
pagliflozin group and the placebo group. The inset
graph shows the data on an expanded y axis. The pri-
mary analysis was performed according to the inten-
tion-to-treat principle and included all the data that
had been obtained up to the end of the planned treat-
ment period for all the patients who had undergone
randomization. The two outcomes were based on the
central blinded adjudication of events reported by the
investigators. For the analysis of the primary outcome,
the assumption of proportional hazards was investi-
gated, and no violations were observed.
and to consider competing risk. The results of
these analyses, which are provided in Table S3,
were similar to the results of the main analyses
reported above.
Discussion
In our trial, the combined risk of cardiovascular
death or hospitalization for heart failure was 25%
lower among the patients who received empa-
gliflozin than among those who received place-
bo, a difference that was primarily related to a
31% lower risk of hospitalization for heart fail-
ure. These benefits were seen in patients receiv-
ing any of the currently recommended drugs for
heart failure, including sacubitril–valsartan, and
were seen regardless of the presence or absence
of diabetes. In addition, empagliflozin was as-
sociated with a lower number of hospitaliza-
tions for heart failure and with a slower rate of
decline in the estimated GFR; the latter effect
was accompanied by a lower risk of serious renal
outcomes.
Our findings with empagliflozin can be com-
pared with the effects of dapagliflozin in the
DAPA-HF trial.5 The current trial was enriched
for patients with a markedly reduced ejection frac-
tion and increased levels of natriuretic peptides,
as compared with the patients in the DAPA-HF
trial (Table S4). Consequently, the incidence of the
primary outcome was approximately 40% higher
in the current trial than in the DAPA-HF trial.
Our trial thus extends the benefits of SGLT2
n engl j med  nejm.org
The New England Journal of Medicine
Downloaded from nejm.org by CARLOS ANDRES TEJEDA PEREZ on October 2, 2020. For personal use only. No other uses without permission.
Copyright © 2020 Massachusetts Medical Society. All rights reserved.
inhibitors to patients with more advanced but
stable heart failure.
In both the current trial and the DAPA-HF
trial, the benefit of the SGLT2 inhibitor on the
primary composite outcome was driven mainly
by a reduction in hospitalizations for heart fail-
ure. The risk of cardiovascular death was 8%
lower with empagliflozin than with placebo in
our trial (hazard ratio, 0.92; 95% CI, 0.75 to 1.12)
and was 18% lower with dapagliflozin in the
DAPA-HF trial (hazard ratio, 0.82; 95% CI, 0.69
to 0.98). Of note, in large-scale trials involving
patients with type 2 diabetes, the risk reductions
in cardiovascular death among patients with
similar cardiovascular histories (i.e., those with
a prior myocardial infarction) were 41% for em-
pagliflozin (hazard ratio, 0.59; 95% CI, 0.44 to
0.79) and 8% for dapagliflozin (hazard ratio,
0.92; 95% CI, 0.69 to 1.23).7,8 Therefore, as noted
by other investigators,1 the effect of SGLT2 in-
hibitors on mortality appears to be heteroge-
neous, with no consistent evidence that one mem-
ber of the drug class is superior to another with
respect to the effects on survival.
In addition to the observed cardiovascular
benefits, empagliflozin slowed the rate of de-
cline in the estimated GFR during double-blind
treatment, and the risk of the composite renal
outcome was lower in the empagliflozin group
than in the placebo group. When measurements
were compared at the start and the end of the
trial after the discontinuation of both empa-
gliflozin and placebo, the estimated GFR declined
more in the placebo group than in the empa-
gliflozin group. These observations are consis-
tent with the benefit observed in trials of SGLT2
inhibitors in patients with type 2 diabetes who
largely did not have heart failure.1 Accordingly,
the ability of empagliflozin to favorably influ-
ence renal function is apparent in patients with
diabetes, in those with heart failure, and in
those with both conditions.
Uncomplicated genital tract infection was
reported more frequently in the empagliflozin
group. The frequency of hypoglycemia, lower
limb amputation, and bone fracture did not dif-
fer between the two groups, even though these
adverse events have been associated with the use
of certain SGLT2 inhibitors in trials involving
patients with type 2 diabetes.9,10 Safety concerns
that have been seen with other drugs for heart
9
 The n e w e ng l a n d j o u r na l of m e dic i n e

Subgroup Empagliflozin Placebo Hazard Ratio (95% CI)

no. of patients with events/total no.
Overall 361/1863 462/1867 0.75 (0.65–0.86)
Baseline diabetes status
Diabetes 200/927 265/929 0.72 (0.60–0.87)
No diabetes 161/936 197/938 0.78 (0.64–0.97)
Age
<65 yr 128/675 193/740 0.71 (0.57–0.89)
≥65 yr 233/1188 269/1127 0.78 (0.66–0.93)
Sex
Male 294/1426 353/1411 0.80 (0.68–0.93)
Female 67/437 109/456 0.59 (0.44–0.80)
Race
White 264/1325 289/1304 0.88 (0.75–1.04)
Black 24/123 48/134 0.46 (0.28–0.75)
Asian 62/337 99/335 0.57 (0.41–0.78)
Other 5/51 14/63 0.41 (0.15–1.14)
Baseline body-mass index
<30 226/1263 322/1300 0.70 (0.59–0.83)
≥30 135/600 140/567 0.85 (0.67–1.08)
Baseline eGFR (CKD-EPI)
≥60 159/969 224/960 0.67 (0.55–0.83)
<60 202/893 237/906 0.83 (0.69–1.00)
HF hospitalization in ≤12 mo
No 208/1286 285/1293 0.71 (0.60–0.85)
Yes 153/577 177/574 0.79 (0.64–0.99)
Cause of heart failure
Ischemic 207/983 236/946 0.82 (0.68–0.99)
Nonischemic 154/880 226/921 0.67 (0.55–0.82)
Baseline NYHA class
II 220/1399 299/1401 0.71 (0.59–0.84)
III or IV 141/464 163/466 0.83 (0.66–1.04)
Heart failure physiology
LVEF ≤30% and NT-proBNP <median 80/699 115/724 0.70 (0.53–0.93)
LVEF ≤30% and NT-proBNP ≥median 169/631 249/661 0.65 (0.53–0.79)
LVEF >30% 108/526 97/475 0.99 (0.76–1.31)
Baseline use of MRA
No 118/557 132/512 0.76 (0.59–0.97)
Yes 243/1306 330/1355 0.75 (0.63–0.88)
Baseline use of ARNi
No 310/1523 369/1480 0.77 (0.66–0.90)
Yes 51/340 93/387 0.64 (0.45–0.89)
0.125 0.25 0.5 1.0 2.0 4.0 8.0

Empagliflozin Better Placebo Better

Figure 2. Primary Outcome in Prespecified Subgroups.

Shown is the risk of the primary outcome in key subgroups of patients. The size of the squares for the hazard ratios is proportional to
the size of the subgroup. The body-mass index is the weight in kilograms divided by the square of the height in meters. Race was report-
ed by the patients. ARNi denotes angiotensin receptor–neprilysin inhibitor, CKD-EPI Chronic Kidney Disease Epidemiology Collabora-
tion, HF heart failure, eGFR estimated glomerular filtration rate, LVEF left ventricular ejection fraction, MRA mineralocorticoid receptor
antagonist, NT-proBNP N-terminal prohormone of brain natriuretic peptide, and NYHA New York Heart Association.

failure (e.g., hypotension, volume depletion, re- Overall, in this trial, empagliflozin was as-
nal dysfunction, bradycardia, and hyperkalemia) sociated with a lower combined risk of cardio-
were not evident with empagliflozin in the cur- vascular death or hospitalization for heart fail-
rent trial. ure than placebo and with a slower progressive

10 n engl j med  nejm.org

The New England Journal of Medicine

Downloaded from nejm.org by CARLOS ANDRES TEJEDA PEREZ on October 2, 2020. For personal use only. No other uses without permission.
Copyright © 2020 Massachusetts Medical Society. All rights reserved.
 Cardiovascular and Renal Outcomes with Empagliflozin

–1
Placebo

Adjusted Mean Change from Baseline –2

in eGFR (ml/min/1.73m2)
–3

Empagliflozin
–4

–5

–6
Between-group difference in slope,
1.73 ml per min per 1.73 m2 per yr;
–7
95% CI, 1.10– 2.37
P<0.001
–8

–9
Base- 4 12 32 52 76 100 124
line Week
No. at Risk
Placebo 1792 1765 1683 1500 1146 745 343 76
Empagliflozin 1799 1782 1720 1554 1166 753 356 80

Figure 3. Changes in the Estimated Glomerular Filtration Rate.

Shown are the adjusted mean changes from baseline in the estimated GFR, as calculated with the Chronic Kidney
Disease Epidemiology Collaboration equation. The I bars indicate the standard error. The on-treatment data were
analyzed with the use of a mixed model for repeated measures that included age and baseline estimated GFR as lin-
ear covariates and sex, region, baseline left ventricular ejection fraction, baseline diabetes status, last projected visit
based on dates of randomization and trial closure, baseline estimated GFR according to visit, and visit according to
treatment interactions as fixed effects. A different model was used to analyze the slope of the change in the estimat-
ed GFR during double-blind treatment, as described in Table 2.

decline in renal function in patients with chron- Supported by Boehringer Ingelheim and Eli Lilly.
Disclosure forms provided by the authors are available with
ic heart failure and a reduced ejection fraction, the full text of this article at NEJM.org.
regardless of the presence or absence of dia- A data sharing statement provided by the authors is available
betes. with the full text of this article at NEJM.org.

Appendix
The authors’ full names and academic degrees are as follows: Milton Packer, M.D., Stefan D. Anker, M.D., Ph.D., Javed Butler, M.D.,
Gerasimos Filippatos, M.D., Stuart J. Pocock, Ph.D., Peter Carson, M.D., James Januzzi, M.D., Subodh Verma, M.D., Ph.D., Hiroyuki
Tsutsui, M.D., Martina Brueckmann, M.D., Waheed Jamal, M.D., Karen Kimura, Ph.D., Janet Schnee, M.D., Cordula Zeller, Dipl.Math.,
Daniel Cotton, M.S., Edimar Bocchi, M.D., Michael Böhm, M.D., Ph.D., Dong‑Ju Choi, M.D., Vijay Chopra, M.D., Eduardo Chuquiure,
M.D., Nadia Giannetti, M.D., Stefan Janssens, M.D., Ph.D., Jian Zhang, M.D., Ph.D., Jose R. Gonzalez Juanatey, M.D., Sanjay Kaul,
M.D., Hans‑Peter Brunner‑La Rocca, M.D., Bela Merkely, M.D., Stephen J. Nicholls, M.D., Sergio Perrone, M.D., Ileana Pina, M.D.,
Piotr Ponikowski, M.D., Naveed Sattar, M.D., Michele Senni, M.D., Marie‑France Seronde, M.D., Jindrich Spinar, M.D., Iain Squire,
M.D., Stefano Taddei, M.D., Christoph Wanner, M.D., and Faiez Zannad, M.D., Ph.D.
The authors’ affiliations are as follows: Baylor Heart and Vascular Institute, Baylor University Medical Center, Dallas (M.P.); Imperial
College (M.P.) and the Department of Medical Statistics, London School of Hygiene and Tropical Medicine (S.J.P.), London, BHF
Glasgow Cardiovascular Research Centre, University of Glasgow, Glasgow (N.S.), and the Department of Cardiovascular Sciences,
University of Leicester and National Institute for Health Research Biomedical Research Centre, Glenfield Hospital, Leicester (I.S.) — all
in the United Kingdom; the Department of Cardiology and Berlin Institute of Health Center for Regenerative Therapies, German Center
for Cardiovascular Research Partner Site Berlin, Charité Universitätsmedizin, Berlin (S.D.A.), Boehringer Ingelheim International, In-
gelheim (M. Brueckmann, W.J.), Faculty of Medicine Mannheim, University of Heidelberg, Mannheim (M. Brueckmann), Boehringer
Ingelheim Pharma, Biberach (C.Z.), Klinik für Innere Medizin III, Saarland University, Homburg–Saar (M. Böhm), and the Department
of Medicine, University Hospital of Würzburg, Würzburg (C.W.) — all in Germany; the Department of Medicine, University of Missis-

n engl j med  nejm.org 11

The New England Journal of Medicine
Downloaded from nejm.org by CARLOS ANDRES TEJEDA PEREZ on October 2, 2020. For personal use only. No other uses without permission.
Copyright © 2020 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

sippi School of Medicine, Jackson (J.B.); National and Kapodistrian University of Athens School of Medicine, Athens University Hospi-
tal Attikon, Athens (G.F.); Washington DC Veterans Affairs Medical Center, Washington DC (P.C.); the Division of Cardiology, Harvard
Medical School and Massachusetts General Hospital, Boston (J.J.); the Division of Cardiac Surgery, St. Michael’s Hospital, University of
Toronto, Toronto (S.V.), Boehringer Ingelheim Canada, Burlington, ON (K.K.), and the Division of Cardiology, McGill University and
Health Centre, Montreal (N.G.) — all in Canada; the Department of Cardiovascular Medicine, Kyushu University, Higashi-ku, Fukuoka,
Japan (H.T.); Boehringer Ingelheim Pharmaceuticals, Ridgefield, CT (J. Schnee, D.C.); Heart Institute, São Paulo University Medical
School, São Paulo (E.B.); the Department of Medicine, Seoul National University, Seoul, South Korea (D.-J.C.); the Department of Car-
diology, Max Super Speciality Hospital, Saket, New Delhi, India (V.C.); the Department of Clinical Cardiology, National Institute of
Cardiology, Mexico City (E.C.); the Department of Cardiology, University Hospital Gasthuisberg of Leuven, Leuven, Belgium (S.J.);
Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College,
Beijing (J.Z.); the Cardiology Department, University Hospital, Santiago de Compostela, Spain (J.R.G.J.); the Department of Cardiology,
Cedars–Sinai Medical Center, Los Angeles (S.K.); Maastricht Heart and Vascular Center, Maastricht, the Netherlands (H.P.B.-L.R.); the
Heart and Vascular Center, Semmelweis University, Budapest, Hungary (B.M.); Victorian Heart Institute and Monash University, Mel-
bourne, VIC, Australia (S.J.N.); Fleni Institute and Hospital El Cruce–Nestor Kirchner, Buenos Aires (S.P.); the Department of Medicine,
Wayne State and Central Michigan Universities, Detroit (I.P.); the Center for Heart Diseases, Wrocław Medical University, Wrocław,
Poland (P.P.); the Cardiovascular Department, Papa Giovanni XXIII Hospital, Bergamo (M.S.), and the Department of Clinical and Ex-
perimental Medicine, University of Pisa, Pisa (S.T.) — both in Italy; the Department of Cardiology, University Hospital Jean Minjoz,
Besançon (M.-F.S.), and Université de Lorraine, INSERM Investigation Network Initiative–Cardiovascular and Renal Clinical Trialists,
Centre Hospitalier Régional Universitaire, Nancy (F.Z.) — both in France; and Internal Cardiology, University Hospital Brno, Brno,
Czech Republic (J. Spinar).

References
1. Zelniker TA, Wiviott SD, Raz I, et al.
SGLT2 inhibitors for primary and second-
ary prevention of cardiovascular and renal
outcomes in type 2 diabetes: a systematic
review and meta-analysis of cardiovascu-
lar outcome trials. Lancet 2019;​393:​31-9.
2. Kato ET, Silverman MG, Mosenzon O,
et al. Effect of dapagliflozin on heart fail-
ure and mortality in type 2 diabetes mel-
litus. Circulation 2019;​139:​2528-36.
3. Packer M. SGLT2 inhibitors produce
cardiorenal benefits by promoting adap-
tive cellular reprogramming to induce a
state of fasting mimicry: a paradigm shift
in understanding their mechanism of ac-
tion. Diabetes Care 2020;​43:​508-11.
4. McMurray JJV, Solomon SD, Inzucchi
SE, et al. Dapagliflozin in patients with
heart failure and reduced ejection frac-
tion. N Engl J Med 2019;​381:​1995-2008.
5. Packer M, Butler J, Filippatos GS, et
al. Evaluation of the effect of sodium-
glucose co-transporter 2 inhibition with
empagliflozin on morbidity and mortality
of patients with chronic heart failure and
a reduced ejection fraction: rationale for
and design of the EMPEROR-Reduced tri-
al. Eur J Heart Fail 2019;​21:​1270-8.
6. Ibrahim NE, Burnett JC Jr, Butler J, et
al. Natriuretic peptides as inclusion crite-
ria in clinical trials: a JACC heart failure
position paper. JACC Heart Fail 2020;​8:​
347-58.
7. Fitchett D, Inzucchi SE, Cannon CP, et
al. Empagliflozin reduced mortality and
hospitalization for heart failure across
the spectrum of cardiovascular risk in the
EMPA-REG OUTCOME trial. Circulation
2019;​139:​1384-95.
8. Furtado RHM, Bonaca MP, Raz I, et
al. Dapagliflozin and cardiovascular out-
comes in patients with type 2 diabetes
mellitus and previous myocardial infarc-
tion. Circulation 2019;​139:​2516-27.
9. Neal B, Perkovic V, Mahaffey KW, et
al. Canagliflozin and cardiovascular and
renal events in type 2 diabetes. N Engl J
Med 2017;​377:​644-57.
10. Donnan JR, Grandy CA, Chibrikov E,
et al. Comparative safety of the sodium
glucose co-transporter 2 (SGLT2) inhibi-
tors: a systematic review and meta-analy-
sis. BMJ Open 2019;​9(1):​e022577.
Copyright © 2020 Massachusetts Medical Society.

12 n engl j med  nejm.org

The New England Journal of Medicine

Downloaded from nejm.org by CARLOS ANDRES TEJEDA PEREZ on October 2, 2020. For personal use only. No other uses without permission.
Copyright © 2020 Massachusetts Medical Society. All rights reserved.